Ship-Based Observations

Water samples were collected by the RV Thetys II within the framework of BOUSSOLE and DyFAMed on March 7 and April 16 2016 at DyFAMed roughly 52 km southeast from Nice, France in the northwestern Mediterranean Sea (Fig. 3.1). Sensor measurements of in situ temperature, salinity, and c(O2), were supplemented by discrete water samples of

c(O2), c(DIC), total alkalinity (AT), and nutrient concentrations (combined nitrite (NO₂-)

and nitrate (NO₃-), silicate (Si(OH)4), and phosphate (PO₄3−)) on March 7 and April 16.

Ship sensor measurements of in situ temperature and salinity were obtained using a Sea-Bird Scientific SBE 9 CTD instrument package. Ship CTD temperature and salinity

measurements were binned vertically every 1 m. The c(O2) were measured using a

Sea-Bird Scientific SBE 43 sensor. Data were visually inspected for erroneous values (e.g. abnormally high or low values) and using statistics, with < 1 % of values considered to be outliers and flagged.

Water samples used to measure c(O2) were obtained on March 7 and April 16 using 12 L

Niskin bottles (General Oceanics 1010X) attached to a carousel water sampler. The bottles were closed at ten depths during the March 7 upcast and at eight depths during the April 16 upcast within the top 1000 m of the water column; four of these depths were within the top 100 m (Fig. 3.2b). Reagents needed for the fixation of oxygen were added to the samples at the time of water sample collection onboard the ship. An automated Winkler titration method with endpoint detection was used after each cruise in the laboratory at the Observatoire Oc´eanologique de Villefranche sur Mer to determine c(O2). Replicates were

obtained to determine instrument precision. c(O2) measured by the rosette-mounted

SBE43 sensor within the top 150 m were at points 15 mmol m−3 higher than the Winkler

measurements (Fig. 3.2b). A trend was present in the depth-dependent ratio between

Winkler measurements and c(O2) Sea-Bird Scientific SBE43 sensor bottle measurements

collected between February 5 and April 16 2016 in the top 100 m. This trend may be related to the methods used to measure Winkler c(O2) measurements (e.g. the storage of

the bottles). Ship Winkler 10 m depth O2 saturation on March 7 was 95.3 %, while ship

SBE43 sensor O2 10 m depth saturation on March 7 was 98.2 % which is closer to the

longterm average of 98.7 % close to the BOUSSOLE region (8◦E to 9◦E, 43◦N to 44◦N)

in March using the World Ocean Atlas (WOA) 2013 climatology

(https://www.nodc.noaa.gov/ cgi-bin/OC5/woa13/woa13oxnu.pl).

Seawater collected at ten depth levels at DyFAMed was also used to measure c(DIC) and

Figure 3.1: (a) The northwestern Mediterranean Sea and glider deployment area (small grey box) superimposed on top of the World Ocean Atlas (WOA) 2013 dissolved oxygen

concentration (c(O2)) March climatology, with accompanying AVISO satellite absolute

mean surface currents (cm s−1, white arrows) for the period of March 6 to April 6, 2016. WOA 2013 climatolgical data were downloaded from the National Oceanography Data

Centre (https://www.nodc.noaa.gov/OC5/woa13/woa13data.html), and AVISO satellite

absolute mean surface currents were downloaded from the Copernicus Marine Environment Monitoring Service (http://marine.copernicus.eu/services-portfolio/access-to-products). (b)

A close-up of the deployment area at the DyFAMed/ BOUSSOLE site. The position

of each glider data point (yellow), the location of the DyFAMed mooring and the approximate location of ship measurements (white), the BOUSSOLE buoy (light blue), and the meteorological buoy (red), are superimposed on top of surface chlorophyll a concentrations (https://oceancolor.gsfc.nasa.gov/products/) on March 24 2016 derived using satellite algorithms described by Hu et al. (2012). As the deployment region is off- shelf, the bathymetry in this area is flat and greater than 2000 m depth.

3.3 Data Collection and Quality 41

Figure 3.2: Calibration of the glider dissolved oxygen concentration (c(O2)) sensor. (a) The

linear regression fit using back-calculated ship pseudo-CalPhase and glider TCPhase for March 7 (orange) and April 16 (blue), along with the corresponding linear equations, r2and root mean squared errors (RMSE). (b) Glider c(O2) measured on March 8-9 and on April

3-4 before (grey), and after (blue) calibration, the ship sensor c(O2) on March 7 (yellow)

and on April 16 (green), and the c(O2) Winkler samples on March 7 (white with red border)

and on April 16 (white with green border). (c) Glider sensor c(O2) at 10 m depth < 10 km

away from the BOUSSOLE buoy (blue spots) are compared with the BOUSSOLE buoy sensor c(O2) measurements (red line) at 10 m depth.

AT. Seawater was transferred into 200 mL borosilicate glass bottles from the Niskin

bottles using tygon tubing. Bottles were rinsed twice, and were allowed to overflow for roughly 20 seconds to ensure the bottle volume was flushed twice. Seawater samples were poisoned, and sealed within glass bottles using greased stoppers. These stoppers were kept in place with elastic bands, and the samples were stored in a dark place (Dickson et al., 2007). A Marianda Versatile INstrument for the Determination of Titration Alkalinity

(VINDTA 3C; www.marianda.com) was used to measure c(DIC) and AT. During this

process, 19 bottles of certified reference material (CRM) supplied by the Scripps Institution of Oceanography (San Diego, CA, USA) were run to calibrate the instrument. Coulometry following standard operating procedure (SOP) 2 was used to measure c(DIC) (Johnson et al., 1985), and potentiometric titration following SOP 3b was used to measure AT (Mintrop et al., 2000). Both SOPs are described in detail by Dickson et al. (2007). The

CO2SYS programme (Van Heuven et al., 2011) was used to derive pHT from c(DIC) and

A_T measurements at DyFAMed for calibrating glider ISFET pHT. The method of using

CO2SYS is explained in more detail in Sect. 3.3.4.

those used for c(O2), c(DIC), and AT. Seawater samples were poisoned with saturated

HgCl2 solution, and were stored in 60 mL polyethylene flasks in a freezer. At the

Observatoire Oc´eanologique de Villefranche sur Mer laboratory, nutrient samples were analysed via a standard automated colourimetry system, using a Seal Analytical continuous flow AutoAnalyser III (AA3). Nitrogen-based nutrients NO₃- and NO₂- were analysed following the procedure described by Bendschneider and Robinson (1952),

Si(OH)4 was analysed following the procedure of Murphy and Riley (1962), and PO₄3−

was analysed according to the procedure by Strickland and Parsons (1972). The detection limits of NO₃- and NO₂-, Si(OH)4, and PO₄3−, were 0.01 mmol m−3, 0.02 mmol m−3, and

0.02 mmol m−3, respectively (de Fommervault et al., 2015).